protocols with qos support 8/10 - 2008 inf5071 – performance in distributed systems
DESCRIPTION
Per-packet QoSTRANSCRIPT
Protocols Protocols with QoS Supportwith QoS Support
8/10 - 2008
INF5071 – Performance in Distributed Systems
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Overview Per-packet QoS
−IP
Per-flow QoS−Resource reservation
QoS Aggregates−DiffServ, MPLS−The basic idea of Network Calculus
Per-packet QoS
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Internet Protocol version 4 (IPv4)
PRE Precedence Field
− Priority of the packet
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Version| IHL |Pre| ToS |0| Total Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Identification |Flags| Fragment Offset | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Time to Live | Protocol | Header Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Destination Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Options | Padding | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
0D T R CPRE
ToSToS Type of Service
D – minimize delay T – maximize throughput R – maximize reliability C – minimize cost
[RFC1349]
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Internet Protocol version 4 (IPv4) 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Version| IHL | DSCP |0 0| Total Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Identification |Flags| Fragment Offset | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Time to Live | Protocol | Header Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Destination Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Options | Padding | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
0 0
Class selector codepointsof the form xxx000
[RFC2474]
DSCP Differentiated Services Codepoint
xxxxx0 reserved for standardizationxxxx11 reserved for local usexxxx01 open for local use, may be
standardized later
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Internet Protocol version 6 (IPv6)
Traffic class − Interpret like IPv4’s DS field
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Version| Traffic Class | Flow Label | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Payload Length | Next Header | Hop Limit | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Source Address + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Destination Address + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Per-flow QoSResource Reservation
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Resource Reservation Reservation is fundamental for reliable enforcement of
QoS guarantees− per-resource data structure (information about all usage)− QoS calculations and resource scheduling may be done based on
the resource usage pattern
− reservation protocols• negotiate desired QoS• transfer information about resource requirements and usage• between the end-systems and all intermediate systems
− reservation operation• calculate necessary amount of resources based on the QoS
specifications• reserve resources according to the calculation (or reject request)
− resource scheduling• enforce resource usage with respect to resource administration
decisions
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Resource Management Phasesuser’s QoS
requirementstime Phase 1:
Phase 2:
Phase 3:
admission test and calculation of QoS guarantees
rejection or renegotiation
resource reservation QoS guarantees to user
negotiation
data transmission QoS enforcement by proper scheduling
monitoring and adaptation “notification”
renegotiation
enforcement
specification
confirmation
renegotiation
stream termination resource deallocation termination
not necessarily an own phase, some protocols start sending at once
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Reservation Directions Sender oriented:
−sender (initiates reservation)• must know target addresses
(participants)• in-scalable• good security
1. reserve
2. reserve
3. reserve
receiver
sender
data flow
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Reservation Directions Receiver oriented:
−receiver (initiates reservation)• needs advertisement before
reservation• must know “flow” addresses
−sender• need not to know receivers• more scalable• in-secure 1. reserve
2. reserve
3. reserve
receiver
sender
data flow
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Reservation Directions Combination?
−start sender oriented reservation
−additional receivers join at routers(receiver based)
receiver
sender
data flow
reserve from nearest router
1. reserve
2. reserve
3. reserve
Per-flow QoSIntegrated Services
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Integrated Services (IntServ) Framework by IETF to provide individualized
QoS guarantees to individual application sessions
Goals:− efficient Internet support for applications which require service
guarantees− fulfill demands of multipoint, real-time applications (like video
conferences) − do not introduce new data transfer protocols
In the Internet, it is based on IP (v4 or v6) and RSVP− RSVP – Resource reSerVation Protocol
Two key features− reserved resources – the routers need to know what resources are
available (both free and reserved)− call setup (admission call) – reserve resources on the whole path from
source to destination
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Integrated Services (IntServ) Admission call:
− traffic characterization and specification• one must specify the traffic one will
transmit on the network (Tspec)• one must specify the requested QoS
(Rspec – reservation specification)
− signaling for setup• send the Tspec and Rspec to all routers
− per-element admission test• each router checks whether the requests
specified in the R/Tspecs can be fulfilled• if YES, accept; reject otherwise
1. request: specify traffic (Tspec), guarantee (Rspec) 1
2
32. consider request against available resources
3. accept or reject
receiver
sender
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Integrated Services (IntServ) IntServ introduces two new services enhancing the
Internet’s traditional best effort:
− guaranteed service• guaranteed bounds on delay and bandwidth• for applications with real-time requirements
− controlled-load service• “a QoS closely to the QoS the same flow would receive from an
unloaded network element” [RFC 2212], i.e., similar to best-effort in networks with limited load
• no quantified guarantees, but packets should arrive with “a very high percentage”
• for applications that can adapt to moderate losses, e.g., real-time multimedia applications
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Integrated Services (IntServ) Both service classes use token bucket to police a packet flow:
− packets need a token to be forwarded
− each router has a b-sized bucket with tokens:if bucket is empty, one must wait
− new tokens are generated at a rate r and added:if bucket is full (little traffic), the token is deleted
− the token generation rate r serves to limit the long term average rate
− the bucket size b serves to limit themaximum burst size
token wait queue
bucket
token generation
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Integrated Services (IntServ) Today implemented
−in every router−for every operating system
(its signaling protocol RSVP is even switched on by default in Windows!)
… and not used
Arguments−too much overhead−too large memory requirements−too inflexible−“net neutrality” argument−no commercial model
QoS AggregatesProtocols
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) IntServ and RSVP provide a framework for per-
flow QoS, but they …−… give complex routers
• much information to handle−… have scalability problems
• set up and maintain per-flow state information• periodically PATH and RESV messages overhead
−… specify only a predefined set of services• new applications may require other flexible services
DiffServ [RFC 2475] tries to be both scalable and flexible
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) ISPs favor DiffServ Basic idea
−multicast is not necessary
−make the core network simple - support to many users−implement more complex control operations at the edge−aggregation of flows –
reservations for a group of flows, not per flowthus, avoid scalability problems on routers with many
flows
−do not specify services or service classes−instead, provide the functional components on which
services can be builtthus, support flexible services
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) Two set of functional elements:
− edge functions: packet classification and traffic conditioning− core function: packet forwarding
At the edge routers, the packets are tagged with a DS-mark (differentiated service mark)− uses the type of service field (IPv4) or the traffic class field
(IPv6)− different service classes (DS-marks) receive different service− subsequent routers treat the packet according to the DS-mark− classification:
• incoming packet is classified (and steered to the appropriate marker function) using the header fields
• the DS-mark is set by marker • once marked, forward classifier marker forward
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) Note, however, that there are no “rules” for classification – it is up
to the network provider
A metric function may be used to limit the packet rate:− the traffic profile may define rate and maximum bursts− if packets arrive too fast, the metric function assigns another marker
function telling the router to delay or drop the packet
classifier marker forwardshaper /dropper
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) In core routers,
DS-marked packets are forwardedaccording to their per-hop behavior (PHB)associated with the DS-tag− the PHB determines how the router resources are used and
shared among the competing service classes− the PHB should be based on the DS-tag only
• no other state in the router− traffic aggregation
• packets with same DS-tag are treated equally• regardless of original source or final destination
− a PHB can result in different service classes receiving different performance
− performance differences must be observable and measurable to be able to monitor the system performance
− no specific mechanism for achieving these behaviors are specified
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
core routers
Differentiated Services (DiffServ)
Edge router:use header fields to lookup right DS-tag and mark packet
Core router:use PHB according to DS-tag to forward packet
fast and scalable due to simple core routers
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Differentiated Services (DiffServ) Currently, two PHBs are under active discussion
− expedited forwarding [RFC 3246]• specifies a minimum departure rate of a class, i.e., a guaranteed
bandwidth• the guarantee is independent of other classes, i.e., enough
resources must be available regardless of competing traffic
− assured forwarding [RFC 2597]• divide traffic into four classes• each class is guaranteed a minimum amount of resources• each class are further partitioned into one of three “drop”
categories(if congestion occur, the router drops packets based on “drop” value)
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Multiprotocol Label Switching (MPLS) Multiprotocol Label Switching
−Separate path determination from hop-by-hop forwarding
−Forwarding is based on labels−Path is determined by choosing labels
Distribution of labels−On application-demand
• LDP – label distribution protocol−By traffic engineering decision
• RSVP-TE – traffic engineering extensions to RSVP
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Multiprotocol Label Switching (MPLS) MPLS works above multiple link layer protocols Carrying the label
−Over ATM• Virtual path identifier or Virtual channel identifier• Maybe shim
−Frame Relay• data link connection identifier (DLCI)• Maybe shim
−Ethernet, TokenRing, …• Shim
Shim?
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Link Layer HeaderShim
Multiprotocol Label Switching (MPLS) Shim: the label itself
Network Layer Header …
Shim
20 bitslabel
3 bitsexperimental
1 bitbottom of stack
8 bits TTL
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Routing using MPLS216.239.51.101
129.42.16.99 80.91.34.111
129.240.148.31
129.240.148.31
66.77.74.20
209.73.164.90192.67.198.54
209.189.226.17
193.99.144.71
81.93.162.20
…Label 12 –
IF 1Label 27 –
IF 2…
Added label
Remove label
Reserved path for this label
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
MPLS Label Stack
ISP 1ISP 1
ISP 2ISP 2
ISP 3
The ISP 1 Classifies the packet Assigns it to a reservation Performs traffic shaping Adds a label to the packet
for routers in his net
The ISP 1 Buys resources from ISP 2The ISP 2 Repeats classifying, assignment,
shaping Adds a label for the routers in his net He pushes a label on the label
stack
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
MPLS Label Stack
ISP 1ISP 1
ISP 2ISP 2
ISP 3
QoS AggregatesNetwork Calculus
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculus Guaranteed Service
− An assured level of bandwidth− A firm end-to-end delay bound− No queuing loss for data flows that conform to a TSpec
TSpec – traffic specification− Describes how customer's traffic must be shaped in the worst case
M
p
rb
tokenbucket
leakybucket
b Double token bucket
(or combined token bucket/leaky bucket)
Token bucket rate rToken bucket depth bPeak rate pMaximum packet size
M
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculus
⎪⎪⎩
⎪⎪⎨⎧
−−
≥+
−−
<+=
rpMbtrtb
rpMbtptM
ta )(arrival curve:
band
widt
h
time M
p
rb
tokenbucket
leakybucket
bM+pt
b+rt
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculusba
ndwi
dth
time M
p
rb
tokenbucket
leakybucket
b
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculusba
ndwi
dth
time
arrival curve service curve
dmax
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculus Using network calculus to scale
Aggregation−Less state in routers
• One state for the aggregate−Share buffers in routers
• Buffer size in routers depends on the TSpec’s rates−Use scheduling to exploit differences in dmax
• Schedule flows with low delay requirements first
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculusba
ndwi
dth
time
Cascaded TSpecSummed TSpec
TSpec(r1,b1,p1,M1)
TSpec(r2,b2,p2,M2)
TSpec(r1+r2,b1+b2,p1+p2,max(M1,M2))
Aggregation
Wastage
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Using Network Calculus
21
11 ppbb + 1
2
221 rpbbb ++
max(M1,M2)
p1+p2
tokenbucket
tokenbucket
r1+r2
leakybucket
r1+p2
Cascaded TSpec: n+1 token buckets
Aggregation
Summary
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Directions of Network QoS Old-style QoS is dead
− ATM,IntServ,DiffServ,Service overlays didn’t take hold
− Causes?• No business case• Bothed standardization• Naïve implementations• No need
Future QoS− Look for fundamental
insights− Develop design principles− Develop analytical tools
• Network calculus
Old-style QoS is dead− X.25 too little, too early− ATM too much, too late− IntServ too much, too early− DiffServ too little, too late− IP QoS not there− MPLS too isolated
QoS through overlays can’t work
Future QoS− Single bit differentiation− Edge-based admission
control− Micropayment
[Crowcroft,Hand,Mortier,Roscoe,Warfield][Liebeherr]
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Directions of Network QoS Old-style QoS is dead
− ATM,IntServ,DiffServ,Service overlays didn’t take hold
− Causes?• No business case• Bothed standardization• Naïve implementations• No need
Future QoS− Look for fundamental
insights− Develop design principles− Develop analytical tools
• Network calculus
Old-style QoS is dead− X.25 too little, too early− ATM too much, too late− IntServ too much, too early− DiffServ too little, too late− IP QoS not there− MPLS too isolated
QoS through overlays can’t work
Future QoS− Single bit differentiation− Edge-based admission
control− Micropayment
[Crowcroft,Hand,Mortier,Roscoe,Warfield][Liebeherr]
Companies do provide QoSAT&T
MPLSEquant
MPLSCable and Wireless
ATMMPLS
TeliaSoneraSDHWDMATM
NortelMPLSSONET/SDHWDM
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Summary Timely access to resources is important for multimedia
application to guarantee QoS – reservation might be necessary
Many protocols have tried to introduce QoS into the Internet, but no protocol has yet won the battle...− often NOT only technological problems, e.g.,
• scalability• flexibility• ...
− but also economical and legacy reasons, e.g.,• IP rules – everything must use IP to be useful• several administrative domains (how to make ISPs agree)• router manufacturers will not take the high costs (in amount of
resources) for per-flow reservations• pricing• ...
INF5071, Carsten Griwodz & Pål HalvorsenUniversity of Oslo
Summary What does it means for performance in
distributed applications?−QoS protocols
• either not present• or used for traffic multiplexes
Applications must adapt to bandwidth competition• either to generic competing traffic• or to traffic within a multiplex
End-to-end QoS can be statistically guaranteed• Overprovisioning in access networks• Network calculus in long-distance networks